Scalable regional genomic surveillance reveals previously unrecognized SARS-CoV-2 recombinant diversity and limited onward transmission

Technical Summary: Scalable Regional Genomic Surveillance of SARS-CoV-2 Recombinants

Problem Statement
Global SARS-CoV-2 surveillance has historically relied on consensus genome sequences, which, while effective for tracking broad viral spread and evolution, offer limited resolution regarding intra-sample diversity. This limitation hinders the reliable detection of recombinant genomes and obscures complex evolutionary processes such as co-infections. Furthermore, as routine surveillance has been scaled back or integrated into broader respiratory monitoring, the tracking of ongoing viral evolution has become more sporadic. There is a critical need for scalable frameworks that can resolve intra-sample diversity and detect recombinant lineages without requiring exhaustive raw-read analysis across entire datasets.

Methodology
The authors analyzed 3,360 high-quality SARS-CoV-2 whole-genome sequences (≥99% coverage, >20-fold depth) generated from regional surveillance in Mecklenburg–Western Pomerania (M-V), Germany, between January 2022 and January 2025. The study employed a two-stage analytical framework designed to balance sensitivity with scalability:

1. Consensus-Based Screening: The REcombination BARcode detector (REBAR) algorithm was applied to consensus sequences to identify candidate recombinant genomes based on lineage-defining mutation barcodes. This step prioritized candidates for further validation.
2. Targeted Raw-Read Validation: To distinguish true recombinant genomes from co-infections, contamination, or consensus-calling artifacts, the authors performed targeted analysis of raw sequencing reads at lineage-defining genomic positions. Classification criteria required that parental variants be supported by at least three consecutive bases with ≥90% read support and a minimum depth of 40. Samples failing these criteria but showing mixed signals were classified as co-infections.
3. Phylogenetic and Spatio-Temporal Analysis: Lineages were assigned using Pangolin and Nextclade. Phylogenetic trees were constructed using RAxML and visualized via Nextstrain/Auspice to reconstruct local spatio-temporal dynamics.
4. Breakpoint Analysis: Recombination breakpoints were mapped and statistically evaluated against a random genomic distribution using permutation tests and Fisher's exact tests (with Benjamini–Hochberg correction) to identify significant hotspots.

Key Results

• Lineage Diversity: The study identified 29 distinct Nextstrain clades and 423 Pangolin lineages. While the regional timeline largely mirrored global trends, specific deviations were observed, including a higher prevalence of Delta variants in early 2022 and a delayed dominance of certain clades compared to global averages.
• Recombinant Detection: Initial screening identified 815 candidate recombinants. After filtering for XBB-derived descendants and applying raw-read validation, 25 sequences were identified as putative true recombinants. Of these, four were assigned to previously described lineages (XCA, XDS, XEC.21, and XBB.1.5.12), leaving 21 sequences representing 17 previously unrecognized, novel recombinant genomes (NRLs). This group of 21 sequences constituted 0.625% of all analyzed sequences.
• Co-infection vs. Recombination: The analysis distinguished between co-infections and recombination. 36 sequences were classified as potential co-infections, predominantly occurring during the Delta-to-Omicron transition (Jan–Mar 2022). Despite frequent co-circulation, only a small subset of recombinants showed evidence of onward transmission.
• Breakpoint Hotspots: Breakpoint analysis revealed non-random clustering in functionally relevant regions. Significant enrichment was found at the ORF1b–spike transition and within the spike (S) gene. While the spike gene showed significant enrichment in circulating recombinant lineages (CRLs), novel recombinants (NRLs) showed a higher proportion of breakpoints in ORF1ab, though this difference was not statistically significant after correction.
• Transmission Dynamics: Most novel recombinants did not establish large transmission chains. Only two NRLs were detected independently on three occasions, suggesting limited local transmission. The study found no evidence that the identified NRLs expanded into major circulating recombinant lineages globally.

Significance and Claims
The paper claims that recombination-aware genomic surveillance complements routine consensus-based monitoring by resolving intra-sample diversity and uncovering a layer of viral diversity that remains hidden to standard lineage assignment frameworks.

The authors assert that their two-stage framework (REBAR screening followed by targeted raw-read validation) provides a practical and scalable strategy for large-scale surveillance. It enables the efficient prioritization of candidate genomes and robust discrimination of true recombination events from co-infections and technical artifacts without the computational burden of analyzing raw reads for every sample.

The study highlights that while recombination is frequently initiated—facilitated by co-circulation and co-infection—the successful epidemiological establishment of recombinant SARS-CoV-2 remains a major constraint. The findings suggest that successful recombinants are likely constrained by immunological factors (e.g., spike protein compatibility with population immunity) and genomic compatibility, rather than just the frequency of co-infection. Ultimately, the work demonstrates that regional, high-resolution surveillance can reveal fine-scale evolutionary dynamics and transient recombinant lineages that are often obscured in global datasets.